I cannot set date after we were rebooted this morning and the clock is off by 5 hours.
Every time ntp tries to update it dies:
time correction of 18001 seconds exceeds sanity limit (1000); set clock manually to the correct UTC time.
Tried:
ntpdate -b 0.us.pool.ntp.org 1.us.pool.ntp.org 2.us.pool.ntp.org 3.us.pool.ntp.org
10 Mar 16:08:42 ntpdate[28568]: step time server 184.105.192.247 offset 18003.254890 sec
when the time should have been 21:08.
If I remove the /etc/localtime link, and report UTC time it is off also, so nothing converting to local.
Please help. I have been at this for over 16hours with no joy.
Gary

What kernel are you using? Check in Linode's Configuration Manager: you should be using latest 3.0.

2.6.18.8-x86_64-linode1

Okay, don't use 2.6.18. You will have better success with syncing your clock if you change to Latest 3.0 and reboot. And make sure you're running ntpd (check your distro's docs for details on how to do this).

I am not in control of the version, and it has been rebooted ~20 times with no problem, so what does the version have to do with it.
I am just trying to figure out what this apparently indiced issue could be and how to resolve it. And NTP is running but dies on every attempt to update. This server has been keeping perfect time with no NTP errors.

That's fine that it has been able to in the past. The thing is, if you stick with outdated software and outdated kernels, you lessen the chance of other people being able to help if something does go wrong...which it finally has for you.

On your prehistoric 2.6.18 kernel, you'll have to set the sysctl xen.independent_wallclock to 1 if you want to be able to set the clock.

It's not like ntpd is "dying". It's intentionally committing suicide because your clock is so wildly wrong it isn't sure the system is sane. If ntpd is passed the -g argument, it will be willing to set the clock no matter how wrong it is, once.

_________________
Matt Nordhoff (aka Peng on IRC)